Use steady_clock instead of high_resolution_clock where monotonic
clock is required. high_resolution_clock is essentially a
system_clock (Wall Clock) therefore may not to be assumed monotonic
since Wall Clock may move backwards due to time/date adjustments.
Fixes issue #638
Signed-off-by: Vlad Zolotarov <vladz@cloudius-systems.com>
This list will store compaction_stats for each ongoing compaction.
That's why register and deregister methods are provided.
This change is important for compaction stats API that needs data
of each ongoing compaction, such as progress, ks, cf, etc.
Signed-off-by: Raphael S. Carvalho <raphaelsc@scylladb.com>
This fixes compile error:
In function `logalloc::segment_zone::segment_zone()':
/home/lmr/Code/scylla/utils/logalloc.cc:412: undefined reference to `logalloc::segment_zone::minimum_size'
collect2: error: ld returned 1 exit status
ninja: build stopped: subcommand failed.
Signed-off-by: Lucas Meneghel Rodrigues <lmr@scylladb.com>
blob_storage defined with attribute packed which makes its alignment
requirement equal 1. This means that its members may be unaligned.
GCC is obviously aware of that and will generate appropriate code
(and not generate ubsan checks). However, there are few places where
members of blob_storage are accessed via pointers, these have to be
wrapped by unaligned_cast<> to let the compiler know that the location
pointed to may be not aligned properly.
Signed-off-by: Paweł Dziepak <pdziepak@scylladb.com>
"This series attempts to make LSA more friendly for large (i.e. bigger
than LSA segment) allocations. It is achieved by introducing segment
zones – large, contiguous areas of segments and using them to allocate
segments instead of calling malloc() directly.
Zones can be shrunk when needed to reclaim memory and segments can be
migrated either to reduce number of zone or to defragment one in order
to be able to shrink it. LSA tries to keep all segments at the lower
addresses and reclaims memory starting from the zones in the highest
parts of the address space."
Originally, lsa allocated each segment independently what could result
in high memory fragmentation. As a result many compaction and eviction
passes may be needed to release a sufficiently big contiguous memory
block.
These problems are solved by introduction of segment zones, contiguous
groups of segments. All segments are allocated from zones and the
algorithm tries to keep the number of zones to a minimum. Moreover,
segments can be migrated between zones or inside a zone in order to deal
with fragmentation inside zone.
Segment zones can be shrunk but cannot grow. Segment pool keeps a tree
containing all zones ordered by their base addresses. This tree is used
only by the memory reclamer. There is also a list of zones that have
at least one free segments that is used during allocation.
Segment allocation doesn't have any preferences which segment (and zone)
to choose. Each zone contains a free list of unused segments. If there
are no zones with free segments a new one is created.
Segment reclamation migrates segments from the zones higher in memory
to the ones at lower addresses. The remaining zones are shrunk until the
requested number of segments is reclaimed.
Signed-off-by: Paweł Dziepak <pdziepak@scylladb.com>
A dynamic bitset implementation that provides functions to search for
both set and cleared bits in both directions.
Signed-off-by: Paweł Dziepak <pdziepak@scylladb.com>
Scattering of blobs from Avi:
This patchset converts the stack to scatter managed_bytes in lsa memory,
allowing large blobs (and collections) to be stored in memtable and cache.
Outside memtable/cache, they are still stored sequentially, but it is assumed
that the number of transient objects is bounded.
The approach taken here is to scatter managed_bytes data in multiple
blob_storage objects, but to linearize them back when accessing (for
example, to merge cells). This allows simple access through the normal
bytes_view. It causes an extra two copies, but copying a megabyte twice
is cheap compared to accessing a megabyte's worth of small cells, so
per-byte throughput is increased.
Testing show that lsa large object space is kept at zero, but throughput
is bad because Scylla easily overwhelms the disk with large blobs; we'll
need Glauber's throttling patches or a really fast disk to see good
throughput with this.
Instead of allocating a single blob_storage, chain multiple blob_storage
objects in a list, each limited not to exceed the allocation_strategy's
max_preferred_allocation_size. This allows lsa to allocate each blob_storage
object as an lsa managed object that can be migrated in memory.
Also provide linearize()/scatter() methods that can be used to temporarily
consolidate the storage into a single blob_storage. This makes the data
contiguous, so we can use a regular bytes_view to examine it.
Our premier allocation_strategy, lsa, prefers to limit allocations below
a tenth of the segment size so they can be moved around; larger allocations
are pinned and can cause memory fragmentation.
Provide an API so that objects can query for this preferred size limit.
For now, lsa is not updated to expose its own limit; this will be done
after the full stack is updated to make use of the limit, or intermediate
steps will not work correctly.
This patch adds a started counter, that is used to mark the number of
operation that were started.
This counter serves two purposes, it is a better indication for when to
sample the data and it is used to indicate how many pending operations
are.
Signed-off-by: Amnon Heiman <amnon@scylladb.com>
boost::heap::binomial_heap allocates helper object in push() and,
therefore, may throw an exception. This shouldn't happen during
compaction.
The solution is to reserve space for this helper object in
segment_descriptor and use a custom allocator with
boost::heap::binomial_heap.
Signed-off-by: Paweł Dziepak <pdziepak@scylladb.com>
LSA memory reclaimer logic assumes that the amount of memory used by LSA
equals: segments_in_use * segment_size. However, LSA is also responsible
for eviction of large objects which do not affect the used segmentcount,
e.g. region with no used segments may still use a lot of memory for
large objects. The solution is to switch from measuring memory in used
segments to used bytes count that includes also large objects.
Signed-off-by: Paweł Dziepak <pdziepak@scylladb.com>
While the objects above max_manage_object_size aren't stored in the
LSA segments they are still considered to be belonging to the LSA
region and are evictable using that region evictor.
Signed-off-by: Paweł Dziepak <pdziepak@scylladb.com>
remove() is the function used to remove every reference to a cf from
the compaction manager. This function works by removing cf from the
queue, and waiting for possible ongoing compaction on cf.
However, a cf may be re-queued by compaction manager task if there
is pending compaction by the end of compaction.
If cf is still referenced by the time remove() returns, we could end
up with an use-after-free. To fix that, a task shouldn't re-queue a
cf if it was asked to stop. The stat pending_tasks was also not
being updated when a cf was removed from the task queue.
Signed-off-by: Raphael S. Carvalho <raphaelsc@scylladb.com>
Add utils::fb_utilities::set_broadcast_address().
Set it to either broadcast_address or listen_address configuration value
if appropriate values are set. If none of the two values above
are set - abort the application.
Signed-off-by: Vlad Zolotarov <vladz@cloudius-systems.com>
New in v2:
- Simplify the utils::fb_utilities::get_broadcast() logic.
From Pawel:
This series enables row cache to serve range queries. In order to achieve
that row cache needs to know whether there are some other partitions in
the specified range that are not cached and need to be read from the sstables.
That information is provied by key_readers, which work very similarly to
mutation_readers, but return only the decorated key of partitions in
range. In case of sstables key_readers is implemented to use partition
index.
Approach like this has the disadvantage of needing to access the disk
even if all partitions in the range are cached. There are (at least) two
solutions ways of dealing with that problem:
- cache partition index - that will also help in all other places where it
is neededed
- add a flag to cache_entry which, when set, indicates that the immediate
successor of the partition is also in the cache. Such flag would be set
by mutation reader and cleared during eviction. It will also allow
newly created mutations from memtable to be moved to cache provided that
both their successors and predecessors are already there.
The key_reader part of this patchsets adds a lot of new code that probably
won't be used in any other place, but the alternative would be to always
interleave reads from cache with reads from sstables and that would be
more heavy on partition index, which isn't cached.
Fixes#185.
The migrator tells lsa how to move an object when it is compacted.
Currently it is a function pointer, which means we must know how to move
the object at compile time. Making it an object allows us to build the
migration function at runtime, making it suitable for runtime-defined types
(such as tuples and user-defined types).
In the future, we may also store the size there for fixed-size types,
reducing lsa overhead.
C++ variable templates would have made this patch smaller, but unfortunately
they are only supported on gcc 5+.
"Fixes: #469
We occasionally generate memtables that are not empty, yet have no
high replay_position set. (Typical case is CL replay, but apparently
there are others).
Moreover, we can do this repeatedly, and thus get caught in the flush
queue ordering restrictions.
Solve this by treating a flush without replay_position as a flush at the
highest running position, i.e. "last" in queue. Note that this will not
affect the actual flush operation, nor CL callbacks, only anyone waiting
for the operation(s) to complete.
To do this, the flush_queue had its restrictions eased, and some introspection
methods added."
This is certainly the right thing to do and seems to fix#403. However
I didn't manage to convince myself that this would cause problems for
binomial_heap, given that binomial_heap::erase() calls siftup()
anyway:
void erase(handle_type handle)
{
node_pointer n = handle.node_;
siftup(n, force_inf());
top_element = n;
pop();
}
void increase (handle_type handle)
{
node_pointer n = handle.node_;
siftup(n, *this);
update_top_element();
sanity_check();
}
As long as we guarantee that the execution order for the post ops are
upheld, we can allow insertion of multiple ops on the same key.
Implemented by adding a ref count to each position.
The restriction then becomes that an added key must either be larger
than any already existing key, _OR_ already exist. In the latter case,
we still know that we have not finished this position and signaled
"upwards".
Previous version dit looping on post execution and signaling of waiters.
This could "race" with an op just finishing if task reordering happened.
This version simplifies the code significantly (and raises the question why
it was not written like this in the first place... Shame on me) by simpy
building a promise-dependency chain between _previous_ queue items and next
instead.
Also, the code now handles propagation of return value from the "Func" pre-op
to the "Post" op, with exceptions automatically handled.
Small utility to order operation->post operation
so that the "post" step is guaranteed to only be run
when all "post"-ops for lower valued keys (T) have been completed
This is a generalized utility mainly to be testable.
